While Large Language Models (LLMs) have exhibited impressive performance in generating long-form content, they frequently present a hazard of producing factual inaccuracies or hallucinations. An effective strategy to mitigate this hazard is to leverage off-the-shelf LLMs to detect hallucinations after the generation. The primary challenge resides in the comprehensive elicitation of the intrinsic knowledge acquired during their pre-training phase. However, existing methods that employ multi-step reasoning chains predominantly fall short of addressing this issue. Moreover, since existing methods for hallucination detection tend to decompose text into isolated statements, they are unable to understand the contextual semantic relations in long-form content. In this paper, we study a novel concept, self-elicitation, to leverage self-generated thoughts derived from prior statements as catalysts to elicit the expression of intrinsic knowledge and understand contextual semantics. We present a framework, SelfElicit, to integrate self-elicitation with graph structures to effectively organize the elicited knowledge and facilitate factual evaluations. Extensive experiments on five datasets in various domains demonstrate the effectiveness of self-elicitation and the superiority of our proposed method.

Recent advances in foundation models have emphasized the need to align pre-trained models with specialized domains using small, curated datasets. Studies on these foundation models underscore the importance of low-data training and fine-tuning. This topic, well-known in natural language processing (NLP), has also gained increasing attention in the emerging field of scientific machine learning (SciML). To address the limitations of low-data training and fine-tuning, we draw inspiration from Heavy-Tailed Self-Regularization (HT-SR) theory, analyzing the shape of empirical spectral densities (ESDs) and revealing an imbalance in training quality across different model layers. To mitigate this issue, we adapt a recently proposed layer-wise learning rate scheduler, TempBalance, which effectively balances training quality across layers and enhances low-data training and fine-tuning for both NLP and SciML tasks. Notably, TempBalance demonstrates increasing performance gains as the amount of available tuning data decreases. Comparative analyses further highlight the effectiveness of TempBalance and its adaptability as an “add-on” method for improving model performance.

This paper describes the system for the Semeval-2022 Task4 ”Patronizing and Condescending Language Detection”.An entity engages in Patronizing and Condescending Language(PCL) when its language use shows a superior attitude towards others or depicts them in a compassionate way. The task contains two parts. The first one is to identify whether the sentence is PCL, and the second one is to categorize PCL. Through experimental verification, the Roberta-based model will be used in our system. Respectively, for subtask 1, that is, to judge whether a sentence is PCL, the method of retraining the model with specific task data is adopted, and the method of splicing [CLS] and the keyword representation of the last three layers as the representation of the sentence; for subtask 2, that is, to judge the PCL type of the sentence, in addition to using the same method as task1, the method of selecting a special loss for Multi-label text classification is applied. We give a clear ablation experiment and give the effect of each method on the final result. Our project ranked 11th out of 79 teams participating in subtask 1 and 6th out of 49 teams participating in subtask 2.